The Research On Video-based Human Upper Limb Motion Analysis Techniques

Video-based human motion analysis is a very important branch of computer vision territory. This system can detect the moving objects from video sequences, and can get the information from the critical moving body. Through the information, analysis system can recognize the human body posture and do further analysis. As an important technology about biomechanics and digital image processing technology, the video-based human motion analysis technique has broad application in human-computer interaction, intelligent surveillance, sports analysis, medical diagnosis, virtual reality and many other fields. Further research is carried based on the background of the Golf auxiliary training system.In this paper, the threshold selection method of canny edge detection has been improved, this new method can be used to select the threshold of video sequences. Based on frame differential method, the method of saving the edge pixels in the dynamic region has been proposed to obtain the edge points of movement. According to the prior knowledge of human body, another algorithm is also proposed to make a further process of the integrated edge points. After that the moving body can be detected more completely.According to the video-based human motion detection techniques and the knowledge and experience of sports, Golf auxiliary training system has been designed and implemented. To process the video of the sportsman lifting up the golf club, the action can be judged right or wrong through this system in the real-time. Further more, it can also show the differences of the arms postures between the measured action and standard action quantitatively.Simulation results show that these proposed methods can not only achieve more accurate and complete detection results but also be more suitable for the videos analysis, which contains background interference and non-rigid problems of human and clothes. When the athletes lifting up the golf club, the motion of the arms can be detected and analyzed completely through the Golf auxiliary training system, the auxiliary training function can be achieved.